In partnership with

World of AI | Edition #45

The Ultimate AI Showdown: ChatGPT vs Gemini vs Perplexity vs Grok

In a comprehensive head-to-head test, four of the most powerful consumer-facing AI chatbots—ChatGPT, Google Gemini, Perplexity, and Grok—were pitted against each other across more than twenty real-world tasks (in a video by ‘Mrwhosetheboss’). These challenges spanned a broad range of everyday needs: translation, critical thinking, image generation, humor, product research, and more. The goal? To answer a practical question that matters to millions of users: Which AI assistant provides the most useful and reliable experience?

Round 1: Basic Reasoning and Visual Recognition

The first few challenges were focused on how well these AI models could reason through a practical spatial problem. For instance, when asked how many large Aerolite suitcases could fit in the trunk of a 2017 Honda Civic, most chatbots produced verbose responses with theoretical possibilities. Grok, on the other hand, offered a simple, confident, and correct answer: two. That kind of confidence and precision gave it an early edge.

In another visual test, the AIs were given a photo of five cooking ingredients, one of which—dehydrated mushrooms—should not have gone into a cake. Only Grok identified the mushrooms correctly and advised against using them, while others misclassified the image as mixed spices, onions, or coffee. This showcased Grok's superior visual analysis in early rounds.

Round 2: Document Generation and Mathematical Reasoning

When tasked with creating a basic Mario Kart scoreboard to track friends' scores, none of the AIs produced an editable, ready-to-use document. Each gave a theoretical template in text form, but the lack of a downloadable file made the entire interaction feel incomplete. A simple spreadsheet would have been faster to create manually.

In math-related tasks, however, performance was better. All four AIs correctly calculated the number of weeks needed to save for a Nintendo Switch 2 ($449 at $42/week), arriving at 11 weeks. They also correctly processed the value of pi multiplied by the speed of light, with minor rounding differences.

Round 3: Language Translation and Nuanced Understanding

Basic translation tasks, such as translating “I’m never gonna give you up” into another language, were handled smoothly by all four platforms. But when asked to translate a sentence filled with homonyms—"I was banking on being able to bank at the bank before visiting the riverbank"—only ChatGPT and Perplexity managed to provide contextually accurate Spanish translations. Grok's translation was too literal and missed the deeper linguistic nuance.

Round 4: Product Research and Real-World Recommendations

Here, the chatbots struggled the most. When asked to recommend red-colored, noise-canceling earbuds under $100, most failed to meet all criteria. Gemini notably hallucinated a Sony WF1000XM6 model that did not exist at the time, while Perplexity became confused and reverted back to a completely unrelated cake recipe from earlier. ChatGPT and Grok performed best in this category but still made occasional missteps with color or product availability.

Adding further complexity, when asked to find earbuds under $10, only ChatGPT, Gemini, and Grok correctly responded that such a product does not realistically exist. Perplexity falsely listed a $40 product as $9.99, misleading the user.

Round 5: Live Web Knowledge and News Awareness

The bots were given a direct link to a product on AliExpress and asked to analyze its content. None of the AIs were able to interpret the link, identify the product, or describe the page details. Instead, they offered general advice, revealing a key weakness in current link-parsing capabilities.

However, when asked about the latest charger release from Ugreen, all four bots correctly acknowledged the announcement of a 500W charger, which had just been released the day before. This reflects improved performance in live knowledge acquisition and real-time awareness, something older generations struggled with.

Round 6: Correlation vs. Causation and Critical Thinking

The AIs were shown a graph correlating cereal consumption to YouTube subscriber growth and asked to draw a conclusion. Gemini and Perplexity recognized that the correlation was purely coincidental and did not imply causation. ChatGPT was less cautious and implied a possible relationship. Grok misunderstood the question entirely and offered the absurd suggestion to eat more cereal to grow a YouTube channel. This test underlined how difficult abstract reasoning and causal inference remain for AI.

Round 7: Generation Capabilities in Text and Visual Media

In text generation, ChatGPT excelled by producing well-structured outputs for use cases like meal planning and YouTube video scripts. Grok impressed with engaging and clickable title suggestions for hypothetical videos, reflecting its training on social media language and trends. Meanwhile, Gemini and Perplexity struggled with prompt adherence in image generation. Inconsistencies ranged from outright denial of capabilities to misinterpreted concepts like “lazy eye.”

Round 8: Memory, Humor, and Deep Research

Memory remains a major weakness across the board. None of the AI systems remembered previous conversation context accurately. When asked how to top a cake that was discussed earlier, none recalled the ingredients or preferences involved.

In humor, Grok stood out by delivering the most natural and timely jokes. Its training on social media platforms like X (formerly Twitter) likely contributed to this result.

When it came to longer-form research tasks, ChatGPT once again led the pack. It provided comprehensive yet focused summaries of recent tech news. Gemini’s answers were lengthy but lacked focus, while Perplexity's results were surface-level.

Round 9: Ecosystem Integration and Overall User Experience

Gemini scored highly for its integrations with Google Workspace tools like Docs, Sheets, Maps, and YouTube, making it especially useful for users embedded in Google’s ecosystem. Grok offered real-time access to trending content on X. ChatGPT allowed the use of custom GPTs tailored for niche interests like competitive Pokémon advice.

Speed was another differentiating factor. ChatGPT and Grok were consistently fast, while Gemini lagged behind, especially on its Pro tier. Perplexity sat in the middle.

Final Scores: Who Comes Out on Top?

After evaluating all 20+ tasks, the final scores were as follows:

• ChatGPT: 29 points

• Grok: 26 points

• Gemini: 24 points

• Perplexity: 22 points

ChatGPT emerged as the most reliable and well-rounded assistant. Its performance was not always flashy, but it consistently delivered accurate and helpful responses across a variety of real-world tasks.

Conclusion: The Best AI Assistant for Most People

Every AI chatbot in this test showed promise in certain areas. Gemini is highly integrated with Google's ecosystem and shows glimpses of power in language tasks. Grok brings personality and boldness, especially in humor and social-context awareness. Perplexity's strength lies in its focus on transparency and sourcing.

But in terms of balance, consistency, and overall usefulness, ChatGPT remains the top choice. It combines solid reasoning, fast response time, excellent generation abilities, and improving integration features. As AI technology continues to evolve, these tools will only get better—but for now, ChatGPT is the most complete offering on the table.

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

1. Join the Superhuman AI newsletter – read by 1M+ people at top companies

2. Master AI tools, tutorials, and news in just 3 minutes a day

3. Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Sign up and start learning AI